Design an A/B Test for Group Video Calls Impact

Quick Overview

This interview question evaluates metric design, causal reasoning, experiment setup, diagnostics, SQL/statistical checks, and recommendations in a realistic interview setting. A strong answer for Design an A/B Test for Group Video Calls Impact states assumptions, handles edge cases, explains trade-offs, and shows how to validate the result clearly.

Design an A/B Test for Group Video Calls Impact

Company: Meta

Role: Data Scientist

Category: Analytics & Experimentation

Difficulty: hard

Interview Round: Technical Screen

##### Scenario Instagram plans an A/B experiment to evaluate the impact of group video calls. ##### Question Design an end-to-end test: hypothesis, randomization unit, sample size and duration. How do you choose the randomization unit given strong network effects? If clustering is infeasible, what alternative designs mitigate interference? Which statistical tests would you apply to continuous versus proportion metrics? How would you present the experiment results to non-technical PMs versus data-science peers? ##### Hints Discuss cluster vs. user-level assignment, geography splits, t-tests vs. z-tests, and storytelling for different audiences.

Quick Answer: This interview question evaluates metric design, causal reasoning, experiment setup, diagnostics, SQL/statistical checks, and recommendations in a realistic interview setting. A strong answer for Design an A/B Test for Group Video Calls Impact states assumptions, handles edge cases, explains trade-offs, and shows how to validate the result clearly.

|Home/Analytics & Experimentation/Meta
Meta logo
Meta
Aug 4, 2025, 10:55 AM
hardData ScientistTechnical ScreenAnalytics & Experimentation
2
0

Design an A/B Test for Group Video Calls Impact

A/B Experiment Design: Group Video Calls on Instagram

Scenario

Instagram wants to evaluate the impact of launching group video calls.

Task

Design an end-to-end experiment that accounts for strong network effects:

  1. State a clear hypothesis, success metrics, and guardrails.
  2. Choose the randomization unit (user, cluster, geography), explaining trade-offs under network interference.
  3. Propose a sample size and duration plan (include formulas and a small numeric example). Account for clustering via design effects.
  4. If clustering is infeasible, describe alternative designs to mitigate interference (e.g., two-stage/saturation, geo/switchback, encouragement designs).
  5. Specify which statistical tests you would use for:
    • Continuous metrics (e.g., time spent, calls per user).
    • Proportion metrics (e.g., % of users who made any group call). Include differences under user-level vs. cluster/geo designs.
  6. Outline how you would present results to non-technical PMs vs. data-science peers.

Hints

  • Discuss cluster vs. user-level assignment, geography splits, and how to measure/limit cross-arm contamination.
  • Contrast t-tests and z-tests; note cluster-robust methods and randomization inference for geo/cluster designs.
  • Tailor the communication: decision and business impact for PMs; assumptions, diagnostics, and methodology for DS peers.

Constraints & Assumptions

  • Preserve the scope, facts, inputs, and requested outputs from the prompt above.
  • If the prompt leaves a detail unspecified, state a reasonable assumption before relying on it.
  • Keep the answer interview-ready: concise enough to present, but concrete enough to implement or evaluate.

Clarifying Questions to Ask Guidance

  • Clarify the business objective, unit of analysis, time window, exposure definition, and primary metric.
  • State assumptions about instrumentation, randomization, sample size, and data quality.
  • Separate descriptive analysis from causal claims.

What a Strong Answer Covers Guidance

  • A metric framework with primary, guardrail, and diagnostic metrics.
  • A credible analysis or experiment design with clear assumptions and bias checks.
  • SQL/statistical logic for segmentation, variance, confidence, and data validation where relevant.
  • An actionable recommendation that explains trade-offs and next steps.

Follow-up Questions Guidance

  • What sanity checks would you run before trusting the result?
  • How would you handle novelty effects, seasonality, or selection bias?
  • What decision would you make if metrics disagree?
Loading comments...

Browse More Questions

More Analytics & Experimentation•More Meta•More Data Scientist•Meta Data Scientist•Meta Analytics & Experimentation•Data Scientist Analytics & Experimentation